XTcalc: MOSFIRE Exposure Time Calculator v2.3

Transcription

1 XTcalc: MOSFIRE Exposure Time Calculator v2.3 by Gwen C. Rudie July 2, Installation using IDL Virtual Machine This is the default way to run the code. It does not require an IDL license. 1.1 Installation First, the user should download and instal the IDL virtual machine software. The software can be downloaded from Note that XTcalc was developed to be launched with idl from the command line. To run IDL from the command line, you must source the idl setup file which is included in the bin directory of the IDL installation. Once you have IDL set up, unpack the tarball and then set the environment variables with the location of the unpacked tarball: MOSFIRE XTCALC with the path where you put the code: REQUIRED. Note the directory will be [local path]/xtcalc dir MOSFIRE WD optional path for input and output files for MOSFIRE. If not set, defaults to MOS- FIRE XTCALC 1.2 Running the code From the command line, in the directory XTcalc dir, run the following line: idl -vm=run XTcalc.sav You can also supply the full path for run XTcalc.sav. 2 Installation with full IDL and license If you already have a working IDL installation, the virtual machine is included. You can run the code as specified above. If you prefer, you can also compile the code and run it in the normal way with IDL as outlined below. 2.1 Installation To run the code using a full IDL installation, XTcalc requires the astronomy idl library ASTROLIB. Note that when running the code outside the virtual machine, any settings in your IDL setup file as well as any previous code run in your IDL session may affect the performance of the code. If you run into issues, please attempt to use the code with the virtual machine, as outlined above. 1

2 The following code and directories are required to run XTcalc. In the bin directory in the packaged version of XTcalc, there are three.pro files which must be compiled prior to running the code. XTcalc.pro: This is the main program which does the actual S/N and exposure time calculations. run XTcalc.pro: This program creates and initializes the GUI. events XTcalc.pro: These programs handle the GUI actions and events. In addition to the code, the following directories are also required: mosfire: This directory contains the filter curves for MOSFIRE. The filter curves are named mosfire [band].txt where [band] in Y, J, H, or K. Mauna Kea sky: This directory contains idl save files which catalog the atmospheric transparency spectra. These files are called mktrans zm [water vapor] [airmass].sav MosfireSkySpec: This directory contains the background sky spectra. They are formatted as idl save files called [band]sky cal pa.sav MosfireSpecEff: This directory contains the throughput spectra for each band for MOSFIRE. The files are called [band]eff.sm.dat The following environment variables should also be set: MOSFIRE XTCALC with the path where you put the directories: REQUIRED MOSFIRE WD optional path for input and output files for MOSFIRE. If not set, defaults to MOS- FIRE XTCALC 2.2 Running the code start IDL.r run XTcalc.r XTcalc.r events XTcalc run XTcalc 3 Using XTcalc XTcalc can be used in two principle modes: Use Line Flux or Use Magnitude. The line flux mode is useful for understanding the S/N within the FWHM of an individual emission line. The magnitude mode calculates the S/N over the full band of observations. In this mode, the default spectral shape if a flat f ν spectrum. However, users can also input their own ASCII spectrum to be considered. The program then scales the chosen spectral shape to match the input magnitude over the band pass considered. 2

3 Figure 1 The XTcalc GUI. The red numerals mark the positions of various features as outlined in the text. 3

4 3.1 Input Please refer to Figure 1 in the following section where the features of XTcalc are described. Sections 1-8 of the GUI allow users to input the properties of their objects. Sections 9-13 pertain to the output of the code. The atmospheric band (Y, J, H, or K) of the observations can be chosen with a pull down menu at position 1. At position 2, the slit width, angular extent of the object, number of exposures, and number of reads (fowler sampling) can be input. Position 3 allows users to choose either the line flux or continuum mode. Section 4 of the GUI pertains to the Use line flux mode. Within this section, the line flux, central wavelength, redshift, and FWHM of the line can be input. Section 5 pertains to the Use Magnitude function. Here the magnitude in AB or Vega can be specified. The spectral shape can be selected using the pull down menu at the top right-hand corner of the box. If the user would like to import their own spectra, choose My Own Spectrum from the pull down menu. You will be prompted to select a file with the input ASCII spectrum with a format (lambda f ν ). The wavelengths can be provided in micron or Å. The spectrum can also be a rest wavelength spectrum, and the observed redshift can be entered. The box labeled 6 is where the user chooses to calculate the S/N given the exposure time or vice versa. Box 7 allows for adjustment in the atmospheric transparency model used in the calculation. Here the user can specify various airmasses and water vapor columns. Position 8 marks the location of the Calculate button which runs XTcalc with the specified parameters and the Exit button which closes the GUI. 3.2 Output Section 9 outputs the calculated values for the signal, various sources of noise, the S/N and the exposure time. The bottom line of the table is useful for checking the saturation level of the data. If the box turns yellow, the exposure time and number of exposures suggest that some of the pixels will pass the 1% linearity. Orange represents the 5% linearity. And red means some of the pixels are saturated. Section 12 provides graphical information regarding the science observations. Section 10 controls what is plotted in the graphical output at position 12. One can choose to plot the Signal and sky residual spectra. Alternatively, users can view the signal, noise and S/N of the spectra. Section 11 allows users to change the plotted range in section 12. Section 13 allows users to print out ASCII spectra of the throughput, transmission, background, signal, noise, and S/N for their proposed observations. 4 Throughput The throughput of the instrument has been measured on sky and is implemented in XTcalc using these measurements. These throughput curves are data taken with MOSFIRE from which the atmospheric absorption has been removed. 4

5 5 Sky Background The background sky spectrum assumed by XTcalc was taken with MOSFIRE during the engineering runs. The spectrum is integrated over the angular extent of the object θ and is scaled considering the slit width selected. Note that when the calculator is used with a broad band magnitude, the angular size of the object is assumed to be the same as (or less than) the slit width. Slit losses are not currently calculated. We will refer to the spectrum of the sky background in units of photons sec 1 (spectral pixel) 1 as β sky. 6 Atmospheric Transparency The atmospheric transparency used in the calculator are the transparency spectra from Gemini Observatory: Lord, S. D., 1992, NASA Technical Memorandum The default parameters are 1.6 mm water vapor column and and airmass of 1, but they can be adjusted in the lower left hand panel of the GUI. 7 The Source Spectrum 7.1 Line Flux The quantity calculated is the signal to noise ratio in the line. If you input a S/N and ask for an exposure time, the program assumes you have input the desired S/N per FWHM of the line. Note that this quantity depends on both the spectral resolution and the entered intrinsic line width. 7.2 Broad Band Magnitude The quantity calculated is the signal to noise ratio per spectral pixel. If you input a S/N and ask for an exposure time, the program assumes you have input the desired S/N per spectral pixel. If you are using the broad band magnitude calculator, the spectrum chosen is read in. The default is a spectrum that is flat in f ν. The spectrum is convolved with the MOSFIRE resolution and resampled onto the MOSFIRE pixel scale. It is multiplied by the filter transmission spectrum, and then the spectrum is normalized such that the flux in the band pass matches the specified broad band magnitude. Multiplied by the atmospheric transparency spectrum and the instrument+telescope throughput spectrum. The resulting spectrum has units of photons sec 1 (spectral pixel) 1 and will hence forth be refereed to as S. 8 Noise Calculation The number of pixels over which the source is spread in the spatial and spectral dimension will factor into the calculation of the noise. The number of pixels along the spatial dimension, N Spatial, is taken to be the spatial extent of the object θ, divided by the spatial pixel scale, 0.18 arcseconds pixel 1. Here, we produce a spectrum of the noise. The units will be photons per spectral pixel. The increase in noise from spreading the light over the spatial direction will have already been factored in. The contributions to the noise, η tot, considered are the read noise and the Poisson noise from the background, dark current (assumed to be electrons s 1 pixel 1 ), and source. 8.1 Single Exposure The read noise, η read, is the detector read noise, RN, in electron pixel 1 read 1 reduced by the Fowler sampling with N reads and increased by the number of pixels. η read = RN Nreads Nspatial (1) 5

6 Table 1. Assumed Values Name Value Units Pixel Scale (spatial) 0.18 arcsec/pixel Pixel Scale (dispersion) 0.24 arcsec/pixel Detector Read Noise 15 electrons/pixel CDS Dark Current electrons/sec/pixel K-band Dispersion Å/pixel H-band Dispersion Å/pixel J-band Dispersion Å/pixel Y-band Dispersion Å/pixel K-band Rθ Product 2534 arcsec H-band Rθ Product 2562 arcsec J-band Rθ Product 2317 arcsec Y-band Rθ Product 2366 arcsec The noise resulting from the dark current is Poisson in nature: η dark (t) = N Spatial Dt (2) Then the total noise per spectral pixel is: η tot = St + β sky t + ηdark 2 (t) + η2 read (3) where D is the dark current in units of electron sec 1 pix 1 (assumed to be 0.005). The source and sky Poisson noise is not multiplied by the number of spatial pixels in this expression because that has already been considered in the calculation of their spectrum in photon sec 1 (spectral pixel) 1. The signal to noise ratio per spectral pixel is then: S/N = S η (4) 8.2 Multiple Exposures with a Two Point Dither If the number of exposures, N exp, is > 1, then a two point dither pattern is assumed in the S/N calculation. The exposure time input is considered to be the total exposure time, and the exposure time of the individual frames is taken to be t/n exp. The total noise per spectral pixel after the frames are differences and co-added is: η tot = N exp S t ( ) t + 2 β sky + ηdark 2 N exp N (t/n exp) + ηread 2 (5) exp The N exp results from the fact that the error in individual exposures is added in quadrature. And the 2 results from the additional contribution of read noise and Poisson noise from the background and dark current which occurs when using difference frames due to the noise in the estimate in the background. This expression simplifies to: η tot = St + 2 (β sky t + ηdark 2 (t) + N expηread 2 ) (6) The S/N per spectral pixel is then simply: S/N = St St + 2 (βsky t + η 2 dark (t) + N expη 2 read ) (7) 6

7 8.3 S/N per resolution element or per FWHM To instead have the S/N per resolution element or the S/N in a line, we multiply by the square root of the number of spectral pixels per resolution element or per FWHM, N Spectral. for the S/N in a line, or N Spectral = N Spectral = FWHM dispersion, (8) Resolution Element, (9) dispersion for a continuum source. The signal to noise reported for the Use line flux option is the signal to noise per FWHM within the line. The signal to noise reported for the Use Magnitude option in the median S/N per spectral pixel where the median is taken over the full band pass of the selected filter. 9 GUI Table The max e- per pixel will turn yellow if the observations will have a pixel past the 1% linearity point of the detector. If the observations surpass the 5% linearity level, the row will turn orange. If the pixel would saturate, the row will turn red. 7